This invention relates to a luminance signal/color signal separation circuit, and more particularly to a circuit for separating Y (luminance) signals and C (carrier chrominance) signals from a composite image signal, to extract them, e.g., in VTR, etc.
In a conventional Y/C separation circuit, a composite image signal (color bar signal) incoming to the input terminal is changed to a 1H delayed signal by a 1H delay circuit. The 1H delayed signal is subtracted from the composite image signal by a subtracter. The signal thus obtained further goes through a bandpass filter and a 1/2 amplifier and is then outputted from the C output terminal as a C signal. On the other hand, the composite image signal is delayed by .DELTA.t corresponding to the delay time of the bandpass filter at a .DELTA.t delay circuit. The C signal is subtracted from the .DELTA.t delayed composite image signal at a subtracter and is then outputted from the Y output terminal as a Y signal.
However, such an image signal processing circuit has the problems that a thin color portion called a half-tone portion occurs in the C signal, whereby the color at the vertical transition portion of a pictorial image is thinned or color fringe is shifted downward in the displayed picture, and that dot crawl due to C signal crosstalk occurs in the Y signal, resulting in considerably degraded picture quality. A further problem with the image signal processing circuit is that vertical resolution lowers in the reproduction of fine character, etc. and thus shading in which color is changed to gray is produced, lacking clearness.
The truth table of the comb filter in the conventional circuit is shown in Table 1. In this Table, asterisk indicates the half tone or the color shift of C signal (dot crawl in the case of Y signal), and double asterisk indicates that a signal is considered as C signal as a whole.
TABLE 1 ______________________________________ a b c ______________________________________ 0 0 0 1 0 1/2* 0 1 -1/2* 1 1 0 1 -1 1** -1 1 -1** ______________________________________
Accordingly, the Applicant has presented a circuit by Japanese Patent Application No. 140921/1987 as shown in FIG. 1. The luminance signal/color signal separation circuit comprises a color signal separation circuit for separating a composite image signal to obtain a first color signal partially including a luminance signal component, a first logic circuit for obtaining a second color signal including a color separation error signal by eliminating the luminance signal component from the first color signal, a second logic circuit for extracting a signal of the second highest potential from the first color signal, the second color signal, and a reference potential to obtain a third color signal exclusing the color separation error signal, and a luminance signal separation circuit for obtaining a luminance signal from the composite signal and the third color signal.
In FIG. 1, a first logic circuit 10 is inputted with a C signal a partially including Y signal component and a signal b being delayed by 1H to output a color separation error signal c' as described later. A second logic circuit 11 is inputted with the C signal a partially including Y signal component and the output signal c' from the first logic circuit 10 to output a C signal c from which the Y signal component is reduced.
A composite image signal inputted to the terminal 1 is changed to a C signal a (partially including Y signal component) (FIG. 2A) at a bandpass filter 12. The C signal a is delivered to a high potential detection circuit (which will be referred to as "MAX" hereinafter) 13 and a low potential detection circuit (which will be referred to as "MIN" hereinafter) 14 of the first logic circuit 10, and, is subjected to 1H delay at a 1H delay circuit 2, resulting in a signal b (FIG. 2B). The signal b is delivered to the MAX 13 and the MIN 14 but a polarity of which is inverted by an inverter (not shown).
The output of the MAX 13 is delivered to a MAX 15, at which it is compared with 0V. On the other hand, the output of the MIN 14 is delivered to a MIN 16, at which it is compared with 0V. The MAX 15 is constituted wherein when the output of the MAX 13 is above 0V, the output of MAX 13 becomes the output. Further, the MIN 16 is constituted wherein when the output of the MIN 14 is below 0V, the output of MIN 14 becomes the output.
The output of the MAX 15 and the output of the MIN 16 are added at an adder 17, resulting in a signal c' (FIG. 2C). The relationship in respect of the outputs of the MAX 15 and the MIN 16, and the signal c' is shown in Table 2. It is to be noted that the combinations indicated by asterisks in this table do not occur actually.
TABLE 2 ______________________________________ MAX 15 MIN 16 c' ______________________________________ 0 0 0 1 0 1 0 -1 -1 1 -1 0 1 1 -1 -1 ______________________________________
The truth table of the first logic circuit 10 is shown in Table 3. As apparent from this Table, in the case that signals a and b have the same level, an input signal is considered as the Y signal to output 0, while in the case of asterisk except for that in Table 2, all input signals are considered as C signals (C signals including color separation error signals).
TABLE 3 ______________________________________ a b c' ______________________________________ 0 0 0 1 0 1* 0 1 -1* 1 1 0 1 -1 1* -1 1 -1* ______________________________________
The signal c' and the signal a are delivered to a MAX 18 which has the same configuration as that of the MAX 13 and performs the same operation as that of the MAX 13, and, on the other hand, are delivered to a MIN 19 which has the same configuration as that of the MIN 14 and performs the same operation as that of the MIN 14. The output of the MAX 18 is delivered to a MIN 20 which has the same configuration as that of the MIN 16 and performs the same operation as that of the MIN 16, at which it is compared with 0V. On the other hand, the output of MIN 19 is delivered to a MAX 21 which has the same configuration as that of the MAX 15 and performs the same operation as that of the MAX 15. Thus, the output of the MIN 20 and the output of the MAX 21 are added at an adder 22, from which the output thus added is taken out in a terminal C as a correct C signal c (FIG. 2D). By subtracting the C signal c from the output signal of a .DELTA.t delay circuit 23 having a delay equal to the delay of the BPF 12 at a subtracter 8, a Y signal e (FIG. 2E) is outputted to a terminal 9.
Data which appear to be C signals, and which are indicated by asterisk in Table 3 are all considered provisionally as C signal at the first logic circuit 10. The second logic circuit 11 carries out error correction of such data.
It is generally known that where the C signal is obtained with a comb filter, if this C signal is completely correct, then both the level and the phase thereof are in correspondence with those of the current line signal a, while if not correct, they are not both in correspondence with them (signals to which asterisk is attached). The second logic circuit 11 obtains a correct C signal by making use of the characteristic mentioned above.
Namely, where the current line signal a and the signal c' are in phase with each other, when a.gtoreq.c', the signal c becomes the signal c', while when a&lt;c, the signal c' becomes the signal a. On the other hand, where the current line signal a and the signal c' are opposite in phase, signals c are all 0. In this instance, "inphase" implies that at least signal c' is considered as signal C, and that "opposite phase" implies that it is considered as a Y signal in the comb filter of the two line system having been described as the prior art. Thus, it cannot be said that the signal c' is a completely correct C signal when it is above the signal a in spite of being in-phase. In such a case, the second logic circuit 11 outputs the signal C with it having an amplitude suppressed to that of the signal a, thus to correct an error produced at the first logic circuit 10.
The truth table indicating the operation of the second logic circuit 11 described above is shown in Table 4, and C signal and Y signal extracted via the first and second logic circuits 10 and 11 are shown in Table 5. Namely, the second logic circuit 11 outputs the second highest potential of the signals a and c' and the reference potential. In Table 5, half tone of C signal or dot crawl of Y signal as in the conventional example (asterisk in Table 1) does not occur at the portion indicated by asterisk, and color blurring at a boundary between subsequent lines having different colors to each others or dot crawl of the Y signal as in the prior art does not occur at the portion indicated by double asterisk.
TABLE 4 ______________________________________ a c' c ______________________________________ 0 0 0 1 0 0 0 1 0 1 1 1 1 -1 0 -1 1 0 ______________________________________
TABLE 5 ______________________________________ a b c y ______________________________________ 0 0 0 0 1 0 1 0* 0 1 0 0** 1 1 0 1 1 -1 1 0 -1 1 -1 0 ______________________________________
Accordingly, high quality pictorial image free from color shift and/or dot crawl which have been encountered with the prior art can be obtained. In addition, since 0 is output as signal c if not data exists in the current line signal a, there is no degradation of the vertical resolution in the case of reproducing fine characters, thus making it possible to obtain a distinct pictorial image.
However, the circuit shown in FIG. 1 previously proposed by the applicant has a large circuit construction and cost reductions are difficult to achieve.
Moreover, the circuit shown in FIG. 1 has a problem as described below.
In case of pictures having no vertical correlation, the C output signals and Y output signals become similar to signals passed through a bandpass filter and a band limited filter, respectively. In these cases, the band widths for C signal and Y signal are determined by the band width of the bandpass filter 12. In contrast, in the case of pictures having vertical correlation, the C output signals and Y output signals become signals which are obtained by passing through a comb filter, and both the band widths are determined by the band width of the bandpass filter 12. Thus, whether the pictures have vertical correlations or not, the band widths of C signal and Y signal which are to be separated are dependent on the band width of the bandpass filter 12.
In order to solve this problem, if the band width for obtaining the C signal of the bandpass filter is made wide, for pictures such as color bar signal, having vertical correlation, cross-luminance which is a phenomenon generating dots at the color borderlines of color bars will not occur, and the frequency characteristics will be improved. However, for random image pictures having no vertical correlation, such as outdoor scenery, there may occur a problem of shading. In contrast, if the band width of the bandpass filter is designed to be relatively narrow, for the above-mentioned picture having no vertical correlation, shading will not occur, but for the above-mentioned picture having vertical correlation, cross-luminance will occur and the frequency characteristics will be deteriorated.
Accordingly, the object of the invention is to provide a luminance signal/color signal separation circuit having a good separation quality with a simplified construction.
According to one aspect of the invention, there is provided a luminance signal/color signal separation circuit comprising: bandpass filtering means for separating a first color signal partially including a luminance signal component from a composite image signal; first delay means for outputting a delayed first color signal as a second color signal by a time period which is multiple of integer of a horizontal scanning period; logic means for outputting a luminance component signal from the first color signal and the second color signal; first separating means for separating a third color signal which is free from the luminance signal component by adding/subtracting the luminance component signal to/from the first color signal; and second separating means for separating luminance signal by adding/subtracting the third color signal to/from the composite image signal.
According to another aspect of the invention, there is provided a luminance signal/color signal separation circuit comprising: wide range bandpass filtering means for separating a first color signal from a composite image signal; delay means for outputting a delayed color signal as a second color signal by a time period which is multiple of integer of a horizontal scanning period; logic means for outputting the color signal with the wide range for pictures having vertical correlation and with a reduced level for pictures having no vertical correlation; and separating means for separating luminance signal by adding/subtracting the color signal to/from the composite image signal.
Other objects and features of this invention will be described with reference to the attached drawings.